Abstract
Background
Hip fractures in the elderly are associated with substantial morbidity and mortality. We compared perioperative outcomes between an internal medicine-led multidisciplinary co-management (IM) model and the traditional orthopedic-led consultation (OC) model in elderly hip fracture patients.
Methods
This retrospective cohort study was conducted at Beijing Jishuitan Hospital Guizhou Hospital from March 2023 to June 2025, enrolling 1095 patients aged ≥65 years with hip fractures (IM, n=872; OC, n=223). Propensity score matching (2:1 optimal maximum matching; caliper 0.45 on the logit of the propensity score) balanced 13 covariates, with exact matching for sex, fracture type, and renal insufficiency. Multivariate regression evaluated associations between the care model and perioperative outcomes.
Results
After matching, all 223 OC patients were successfully matched (final cohort 668: 445 IM, 223 OC), with all covariates achieving SMD < 0.1 (maximum 0.078). The IM group showed shorter hospital stay (10.0 vs 14.3 days), shorter time to surgery (42.3 vs 110.1 hours), less blood loss, and lower costs (all p < 0.001). Surgery within 48 hours was more frequent (70.3% vs 8.5%, p < 0.001), and allogeneic transfusion (49.4% vs 68.2%, p < 0.001) and one-year mortality (7.0% vs 12.6%, p = 0.016) were lower in the IM group. Multivariate analysis confirmed an independent association with lower mortality (OR 0.492, 95% CI 0.280–0.863, p = 0.013).
Conclusion
Internal medicine-led multidisciplinary co-management may be associated with shorter preoperative waiting time and hospital stay, reduced blood loss, allogeneic transfusion, one-year mortality, and medical costs. This model may serve as a reference strategy for perioperative management of geriatric hip fracture where multidisciplinary internal medicine resources are available; multicenter prospective studies are warranted.
Keywords: geriatric hip fracture, internal medicine-led, multidisciplinary team, propensity score matching, mortality
Background
Hip fracture is a common and devastating injury among the elderly, primarily driven by osteoporosis-induced reduction in bone mineral density.1 Even low-energy trauma such as a slip on level ground can precipitate fracture.2 Due to the high prevalence of pre-existing comorbidities, the one-year mortality rate after hip fracture reaches 20–30%, and long-term disability exceeds 50%, earning it the clinical designation of “the last fracture of life”.3 The Global Burden of Disease study indicates that disability-adjusted life years attributable to hip fracture increase exponentially with age, imposing a substantial and growing economic burden on healthcare systems worldwide.2 In China, the population aged 65 years and older has reached approximately 223.65 million, accounting for 15.9% of the total population,4 and the incidence of hip fracture continues to rise with accelerating population aging.
In response, the National Health Commission of China issued the Guidelines for the Diagnosis and Management of Geriatric Hip Fracture (2022 Edition),5 which recommends multidisciplinary team (MDT) management with standardized emergency green channels and assessment protocols. However, no unified regulation was made regarding which specialty should lead the team. Consequently, the prevailing MDT model in China adopts an orthopaedic-led, consultation-based approach,6 wherein medical comorbidities are managed reactively through referrals to other departments. This model lacks proactive, continuous medical oversight and often results in prolonged preoperative delays.
The critical importance of surgical timing is well documented: Bani Essa et al7 demonstrated that delayed surgery significantly increases one-year complications in elderly hip fracture patients. Growing evidence supports the superiority of comanaged or specialty-led models over traditional consultation-based care. Internationally, orthogeriatric comanagement models have demonstrated reductions in mortality, complications, and length of stay.8–10 In China, Liu et al11 reported that orthogeriatric co-management significantly improved outcomes in elderly hip fracture patients. A recent meta-analysis by Yavuz Veizi et al12 further confirmed that geriatric co-management improves functional outcomes and shortens hospitalization. Despite this growing body of evidence, the optimal leadership structure within the MDT, specifically whether internists should serve as the primary attending physicians, remains insufficiently studied. In addition, perioperative blood-management strategies, such as intravenous or topical tranexamic acid and restrictive transfusion thresholds, have been shown to reduce blood loss and allogeneic transfusion in major hip surgery,13 reinforcing the clinical relevance of transfusion-related outcomes when comparing care models.
To address this gap, our center established a Geriatric Hip Fracture Treatment Center in March 2023. We adopted an internal medicine-led MDT model with participation from traumatology/arthroplasty, anesthesiology, rehabilitation medicine, and nutrition departments. The staffing includes specialists in cardiology, respiratory medicine, endocrinology, and neurology, covering the full spectrum of common geriatric comorbidities. After two years of operation, we employed propensity score matching (PSM) to perform a rigorous retrospective comparison between the medicine-led co-management model and the traditional orthopaedic-led consultation model.
Methods
Study Design and Participants
This retrospective cohort study was conducted at Beijing Jishuitan Hospital Guizhou Hospital, a tertiary referral center in Guizhou Province, China. Consecutive patients aged 65 years or older with radiographically confirmed femoral neck, intertrochanteric, or subtrochanteric fractures admitted between March 2023 and June 2025 were retrospectively identified. Subtrochanteric fractures were pooled with intertrochanteric fractures for the analysis; after matching, 12 patients had subtrochanteric fractures included in this pooled category.
Inclusion criteria were: (1) age 65 years or older; (2) radiographically confirmed femoral neck or intertrochanteric/subtrochanteric fracture; (3) injury-to-admission interval of 21 days or less; and (4) complete surgical treatment at our institution with comprehensive clinical follow-up data.
Exclusion criteria included: (1) pathological or periprosthetic fractures; (2) conservative management without surgery; and (3) missing critical clinical or follow-up information.
Patients were allocated to two groups based on the admission pathway. Following the establishment of a dedicated geriatric hip fracture treatment center at our institution, patients admitted to this center received internal medicine-led multidisciplinary co-management (IM group). Patients admitted to conventional orthopedic wards with on-demand internal medicine consultation for comorbidity management constituted the orthopedic-led consultation group (OC group). Assignment depended on the ward of initial admission rather than on clinician or patient preference, and patients were not transferred between pathways after allocation (Figure 1).
Figure 1.
Flowchart of patient selection. PSM: 2:1 deterministic optimal maximum matching; fixed caliper = 0.45 on the logit of the propensity score; exact matching for sex, fracture type, and renal insufficiency; lexicographic tie-breaking.
Treatment Protocols
All surgical procedures were performed by senior attending surgeons or above under standardized protocols. Intertrochanteric fractures were managed primarily with intramedullary nail fixation. Patients with severe osteoarthritis, femoral head necrosis, or Garden type III/IV femoral neck fractures underwent joint arthroplasty. Stable femoral neck fractures could be managed with cannulated screw fixation based on patient preference. All surgeries employed spinal anesthesia.14
In the IM group, a comprehensive multidisciplinary pathway was activated immediately upon admission. Patients underwent simultaneous preoperative laboratory testing, electrocardiography, and imaging studies through a dedicated geriatric fracture green channel. Anesthesiologists performed bedside pain management and ASA classification within hours of arrival. Cardiologists, pulmonologists, endocrinologists, and other internal medicine subspecialists conducted bedside multisystem comorbidity assessment and optimization within 24 hours. Once surgical fitness criteria were met, urgent or emergency surgery was scheduled without delay. Postoperatively, rehabilitation physicians collaborated with surgeons to develop individualized recovery plans, accompanied by nutritional support, standardized anti-osteoporosis therapy, and venous thromboembolism prophylaxis. The clinical pathway is illustrated in Figure 2.
Figure 2.
Clinical pathway for elderly hip fracture patients under the medicine-led co-management model.
In the OC group, orthopedic surgeons retained primary responsibility throughout the treatment course. Internal medicine consultations for comorbidities such as hypertension, cardiovascular disease, and respiratory conditions were arranged on a case-by-case basis without a standardized multidisciplinary assessment framework. Surgical techniques, postoperative rehabilitation, and basic anti-osteoporosis management followed the same standards as the IM group.
Outcome Measures
Primary outcomes included: length of hospital stay, time from admission to surgery, operation duration, intraoperative blood loss, total hospitalization costs, allogeneic transfusion rate, one-year mortality, and the proportion of patients undergoing surgery within 48 hours of admission. Secondary outcomes comprised mobility status at discharge, categorized as independent ambulation, ambulation with assistive devices, or bedridden, and 1-year survival status ascertained through outpatient visits and telephone follow-up.
Comorbidity Ascertainment
Nine comorbidities were identified from discharge diagnosis text: hypertension, diabetes mellitus, coronary heart disease, cerebral infarction, chronic heart failure, chronic obstructive pulmonary disease (COPD), renal insufficiency, pulmonary infection, and venous thrombosis. Two senior residents independently extracted comorbidity information through keyword identification from discharge summaries. Discrepancies were adjudicated by an attending physician or above. Comorbidities were dichotomized (present/absent) for inclusion in the propensity score model.
Statistical Analysis
Statistical analyses were performed using R 4.6.1 (clue) and Python 3 (scipy, statsmodels), with fully deterministic matching routines cross-validated between platforms. Propensity scores were estimated via unconditional logistic regression including 13 covariates: age, sex, fracture type, surgical method, and nine comorbidities described above. We conducted 2:1 matching without replacement using the deterministic optimal maximum matching (Hungarian) algorithm, with a fixed caliper of 0.45 on the propensity-score logit scale. Among solutions with identical pair numbers and total logit distance, the lexicographically smallest control assignment was selected, and partial matches retained. Exact matching was enforced for sex, fracture type, and renal insufficiency. Covariate balance was assessed using standardized mean differences (SMD); SMD < 0.1 indicated adequate balance.15
All covariates were baseline admission characteristics. Pulmonary infection and venous thrombosis were presenting diagnoses, whereas surgical method was determined after admission but before surgery; we acknowledge this variable may be downstream, as noted in the limitations. Continuous variables deviating significantly from normality (Shapiro–Wilk test) were reported as medians (interquartile ranges) [M(IQR)] and compared using the Mann–Whitney U-test. Categorical variables were compared with the χ2-test, with Fisher’s exact test used when expected cell frequencies were <5. After matching, multivariable logistic and linear regression models were constructed for one-year mortality and for hospital stay and total costs, respectively, adjusting for all 13 baseline covariates. Given the marked right skewness of hospital costs, linear regression results were supplemented with nonparametric univariable comparisons and interpreted cautiously. All tests were two-sided with α = 0.05.
E-values were calculated to evaluate the robustness of mortality estimates to unmeasured confounding, representing the minimum risk-ratio association an unmeasured confounder would need with both exposure and outcome to fully explain the observed association conditional on measured covariates16. Given the 2:1 matching without replacement, all between-group analyses were performed in the matched cohort; potential residual nonindependence within matched sets is addressed in the limitations. A prespecified sensitivity analysis adopted a narrower caliper plateau (0.155–0.170 on the propensity-score logit scale), within which matched sets remained invariant.
Ethical Approval
This study was approved by the Institutional Ethics Committee of Beijing Jishuitan Hospital Guizhou Hospital (approval number: KT2026042001); although the clinical care covered by the study period preceded the ethics application, ethical approval was obtained before study-specific data collection was initiated, in line with the retrospective design. The requirement for individual informed consent was waived. Patient data were fully anonymized, and confidentiality was strictly maintained. All procedures were performed in accordance with the Declaration of Helsinki.
Results
Baseline Characteristics and Propensity Score Matching
A total of 1095 patients were included in the initial cohort, comprising 872 patients in the IM group and 223 in the OC group. Following optimal maximum matching, all 223 OC group patients (100%) were successfully matched, yielding 445 IM group patients and a final analytical cohort of 668 patients. The theoretical maximum of 446 internal medicine patients was not reached because matching was performed without replacement: one consultation patient had only a single eligible donor under the caliper and exact-matching criteria, and this donor was allocated to another patient by the globally optimal matching algorithm. Prior to matching, several covariates including age, pulmonary infection, and COPD exhibited SMD values exceeding 0.1, indicating substantial baseline imbalance between groups. After matching, the maximum SMD among all 13 covariates was 0.078 (pulmonary infection), with all values below the 0.1 threshold. Sex, fracture type, and renal insufficiency achieved exact balance (SMD < 0.01) by design. Baseline characteristics before and after matching are presented in Table 1, and the SMD comparison is illustrated in Figure 3. The distribution of propensity scores before and after matching is displayed in Figure 4, demonstrating substantial overlap following the matching procedure.
Table 1.
Baseline Characteristics of Patients Before and After Propensity Score Matching
| Variable | IM (n=872) | OC (n=223) | SMD | IM (n=445) | OC (n=223) | SMD |
|---|---|---|---|---|---|---|
| Before Matching | After Matching | |||||
| Age (years), M(IQR) | 80.0 (74,85) | 79.0 (72,83) | 0.296 | 78.0 (72,83) | 79.0 (72,83) | 0.008 |
| Male, n (%) | 247 (28.3) | 53 (23.8) | 0.104 | 105 (23.6) | 53 (23.8) | 0.004 |
| Femoral neck fracture, n (%) | 491 (56.3) | 135 (60.5) | 0.086 | 270 (60.7) | 135 (60.5) | 0.003 |
| Internal fixation, n (%) | 440 (50.5) | 103 (46.2) | 0.086 | 217 (48.8) | 103 (46.2) | 0.052 |
| Hypertension, n (%) | 458 (52.5) | 123 (55.2) | 0.053 | 236 (53.0) | 123 (55.2) | 0.043 |
| Diabetes mellitus, n (%) | 176 (20.2) | 43 (19.3) | 0.023 | 84 (18.9) | 43 (19.3) | 0.010 |
| Coronary heart disease, n (%) | 27 (3.1) | 10 (4.5) | 0.073 | 15 (3.4) | 10 (4.5) | 0.057 |
| Cerebral infarction, n (%) | 164 (18.8) | 38 (17.0) | 0.046 | 65 (14.6) | 38 (17.0) | 0.067 |
| Chronic heart failure, n (%) | 41 (4.7) | 12 (5.4) | 0.031 | 23 (5.2) | 12 (5.4) | 0.010 |
| COPD, n (%) | 101 (11.6) | 9 (4.0) | 0.284 | 16 (3.6) | 9 (4.0) | 0.023 |
| Renal insufficiency, n (%) | 54 (6.2) | 10 (4.5) | 0.076 | 19 (4.3) | 10 (4.5) | 0.010 |
| Pulmonary infection, n (%) | 399 (45.8) | 69 (30.9) | 0.308 | 154 (34.6) | 69 (30.9) | 0.078 |
| Venous thrombosis, n (%) | 319 (36.6) | 87 (39.0) | 0.050 | 177 (39.8) | 87 (39.0) | 0.016 |
Abbreviations: SMD, standardized mean difference; M(IQR), median (interquartile range); IM, internal medicine-led group; OC, orthopedic-led consultation group; COPD, chronic obstructive pulmonary disease.
Figure 3.
Comparison of standardized mean differences (SMD) of all covariates before and after propensity score matching. Red circles = before matching; blue squares = after matching; vertical dashed line = SMD = 0.1 threshold.
Figure 4.
Histogram distribution of propensity scores before (left) and after (right) matching.
Perioperative Outcomes After Matching
The IM group showed significantly better outcomes across all primary endpoints. Median length of hospital stay was shorter in the IM group (10.0 (8.2–12.2) vs 14.3 (11.3–20.0) days, Z = 12.017, p < 0.001). Time from admission to surgery was substantially reduced (42.3 (27.2–52.4) vs 110.1 (72.2–159.9) hours, Z = 15.018, p < 0.001), and the proportion of patients undergoing surgery within 48 hours was markedly higher in the IM group (70.3% vs 8.5%, χ2 = 227.080, p < 0.001). Operation time did not differ significantly between the two groups (75.0 (60.0–100.0) vs 79.0 (60.0–110.0) minutes, Z = 1.358, p = 0.174) (Table 2, Figure 5).
Table 2.
Comparison of Perioperative Outcomes After Propensity Score Matching
| Outcome | IM Group (n=445) | OC Group (n=223) | Statistic | p value |
|---|---|---|---|---|
| Length of stay (d), M(IQR) | 10.0 (8.2–12.2) | 14.3 (11.3–20.0) | Z = 12.017 | <0.001 |
| Time to surgery (h), M(IQR) | 42.3 (27.2–52.4) | 110.1 (72.2–159.9) | Z = 15.018 | <0.001 |
| Surgery within 48 h, n (%) | 313 (70.3) | 19 (8.5) | χ2 = 227.080 | <0.001 |
| Operation time (min), M(IQR) | 75.0 (60.0–100.0) | 79.0 (60.0–110.0) | Z = 1.358 | 0.174 |
| Intraoperative blood loss (mL), M(IQR) | 100 (50–150) | 200 (100–300) | Z = 10.018 | <0.001 |
| Total cost (CNY), M(IQR) | 34,556.9 (27,808–41,929) | 52,554.2 (41,609–66,700) | Z = 13.226 | <0.001 |
| Allogeneic transfusion, n (%) | 220 (49.4) | 152 (68.2) | χ2 = 21.104 | <0.001 |
| One-year mortality, n (%) | 31 (7.0) | 28 (12.6) | χ2 = 5.764 | 0.016 |
Abbreviations: M(IQR), median (interquartile range); Mann–Whitney U-test for continuous variables; χ2-test for categorical variables; CNY, Chinese Yuan.
Figure 5.
Boxplots of continuous perioperative outcomes after PSM. Top row: length of stay, time to surgery, operation time; bottom row: intraoperative blood loss, total hospitalization cost.
Intraoperative blood loss was reduced in the IM group (100 (50–150) vs 200 (100–300) mL, Z = 10.018, p < 0.001). Total hospitalization costs were substantially lower (34,556.9 (27,808–41,929) vs 52,554.2 (41,609–66,700) CNY, Z = 13.226, p < 0.001). The allogeneic transfusion rate was significantly lower in the IM group (49.4% vs 68.2%, χ2 = 21.104, p < 0.001). The one-year all-cause mortality was significantly reduced in the IM group (7.0% vs 12.6%, χ2 = 5.764, p = 0.016) (Figure 6).
Figure 6.
Bar chart of effect sizes (Cohen’s d/risk difference) for all outcomes. Green bars: p < 0.01; Orange bar: p = 0.016 (0.01 ≤ p < 0.05); grey bars: not significant (operation time, p = 0.174).
Multivariate Analysis
After adjusting for all 13 baseline covariates in the matched cohort, multivariate logistic regression confirmed that the internal medicine-led co-management model was independently associated with reduced one-year mortality (OR = 0.492, 95% CI 0.280–0.863, Wald χ2 = 6.117, p = 0.013; Table 3, Figure 7). Multivariate linear regression showed that the IM model was independently associated with an approximately 6.7-day shorter hospital stay (β = −6.7, 95% CI −7.8 to −5.7, t = −12.553, p < 0.001) and with total hospitalization costs approximately 19,700 CNY lower (β = −19,711.7, 95% CI −22,186.3 to −17,237.1, t = −15.641, p < 0.001; Table 3).
Table 3.
Multivariate Regression Analysis of Care Model Effect on Outcomes After Matching
| Outcome | β/OR | 95% CI | Statistic | p value |
|---|---|---|---|---|
| One-year mortality (Logistic) | OR = 0.492 | 0.280–0.863 | Wald χ2 = 6.117 | 0.013 |
| Length of stay (Linear) | β = −6.7 | −7.8 to −5.7 | t = −12.553 | <0.001 |
| Total cost (Linear) | β = −19,711.7 | −22,186.3 to −17,237.1 | t = −15.641 | <0.001 |
Notes: Models adjusted for all 13 baseline covariates.
Figure 7.
Forest plot of multivariate logistic regression for one-year mortality. Squares = OR; horizontal lines = 95% CI; vertical reference line at OR = 1; *Indicates statistically significant p values (p < 0.05).
The E-value for the observed association was 3.48, indicating that an unmeasured confounder would need to be associated with both the co-management model and one-year mortality by a risk ratio of 3.48-fold each (above and beyond the measured covariates) to fully explain away the observed OR. The corresponding E-value for the upper bound of the 95% CI (0.863) was 1.59, indicating that the upper limit of the confidence interval could be explained away by even weak unmeasured confounding (Figure 8).
Figure 8.
E-value sensitivity analysis for the association between internal medicine-led co-management and one-year mortality. (A) E-value plot: E-value = 3.48 for the point estimate (OR = 0.492) and 1.59 for the upper bound of the 95% CI (0.863). (B) Required confounding strength: an unmeasured confounder would need a risk ratio of at least 3.48 (or 1.59 for the CI upper bound) with both exposure and outcome to explain away the observed association.
A sensitivity analysis using the narrower caliper plateau (0.155–0.170, n=629) yielded consistent results, with a one-year mortality OR of 0.462 (95% CI 0.258–0.827, p = 0.009), length of stay β = −6.9 days (95% CI −8.0 to −5.9), and total cost β = −19,590 CNY (95% CI −22,145 to −17,035), indicating robustness to caliper specification.
Postoperative Mobility
Mobility status at discharge differed significantly between the two groups (χ2 = 12.283, p = 0.002). The IM group had a higher proportion of patients achieving independent ambulation (41.1% vs 31.8%) and a lower proportion of bedridden patients (7.2% vs 14.8%) compared with the OC group (Table 4).
Table 4.
Comparison of Postoperative Mobility at Discharge
| Mobility Level | IM Group (n=445), n (%) | OC Group (n=223), n (%) | p value |
|---|---|---|---|
| Independent ambulation | 183 (41.1) | 71 (31.8) | 0.020 |
| Ambulation with assistive device | 230 (51.7) | 119 (53.4) | 0.682 |
| Bedridden | 32 (7.2) | 33 (14.8) | 0.002 |
Discussion
This propensity score-matched study provides real-world evidence that an internal medicine-led multidisciplinary co-management model is associated with significantly better perioperative outcomes in elderly patients with hip fractures compared with the traditional orthopedic-led consultation approach. After rigorous balancing of 13 baseline covariates, the IM group demonstrated clinically meaningful reductions across all major endpoints, including preoperative waiting time, hospital stay, intraoperative blood loss, transfusion requirements, hospitalization costs, and one-year mortality.
The most plausible mechanism underlying these associations appears to be the acceleration of preoperative assessment and optimization. In the IM model, internal medicine teams conducted comprehensive bedside multisystem evaluations immediately upon admission, concurrent with standard preoperative investigations. This parallel processing eliminated the sequential bottleneck inherent in the consultation-based model, where each comorbidity required a separate referral and waiting period. The result was a substantially higher proportion of patients undergoing surgery within 48 hours, from 8.5% to 70.3%, consistent with previous reports that integrated multidisciplinary pathways can substantially reduce surgical delays.10,11,16,17
The reduction in preoperative waiting time carries important clinical implications. Prolonged immobility before surgery is associated with increased risks of pulmonary infection, venous thromboembolism, pressure ulcers, and functional decline in elderly patients.3,12 By minimizing this vulnerable period, the IM model likely reduced the incidence of these immobility-related complications. Furthermore, early preoperative optimization of comorbidities, including correction of anemia, optimization of cardiac and pulmonary function, and glycemic and blood pressure control, may have contributed to reduced intraoperative stress responses and diminished bleeding, thereby lowering transfusion requirements.18,19
These converging benefits were accompanied by an observed one-year mortality that was approximately 44% lower (12.6% vs 7.0%), an association that persisted after multivariate adjustment (OR 0.492, 95% CI 0.280–0.863, p = 0.013). This magnitude of mortality reduction is consistent with findings from orthogeriatric co-management studies in other healthcare settings20–24 and underscores the potential of multidisciplinary care models to meaningfully impact patient survival. Nevertheless, this association should not be attributed to internal medicine leadership per se, as the dedicated geriatric pathway, expedited multidisciplinary assessment, and emergency green-channel organization introduced simultaneously with the model may also have contributed.
To assess the robustness of this mortality finding against potential unmeasured confounding, we conducted an E-value sensitivity analysis.25 The E-value for the observed association (OR = 0.492) was 3.48, indicating that an unmeasured confounder would need to be associated with both the treatment model and mortality by a risk ratio of at least 3.48 to fully explain away the observed protective effect. The E-value for the upper bound of the 95% confidence interval was 1.59. These results suggest that the point estimate is relatively robust to unmeasured confounding, whereas the upper confidence limit, lying just below 1, remains sensitive to even weak confounding. This is consistent with the comprehensive covariate adjustment and rigorous PSM design employed in this study.
From a health economics perspective, the IM model reduced hospitalization costs by approximately 19,700 CNY per patient (multivariable-adjusted estimate), a reduction attributable to shorter hospital stays, fewer transfusions, and decreased utilization of consumables and complication-related treatments. These findings align with previous reports demonstrating the cost-effectiveness of integrated care models per patient.26 The observed reductions in intraoperative blood loss and allogeneic transfusion are also consistent with growing evidence that structured perioperative blood-management strategies, including tranexamic acid administration and restrictive transfusion thresholds, reduce transfusion requirements in major hip surgery.13
Compared with previous Chinese studies of multidisciplinary hip fracture care, our study offers several distinctive contributions. While most prior Chinese reports have focused on orthopedic-led or geriatrician-led models in major metropolitan teaching hospitals,27–33 our investigation was conducted at a regional tertiary hospital in Guizhou Province, representing a less well-resourced setting that more closely resembles the circumstances of many secondary and tertiary hospitals across China. The favorable outcomes achieved in this setting suggest that the internal medicine-led model is feasible and associated with favorable outcomes even in the absence of a fully developed geriatric medicine department.
The mobility outcomes observed in this study warrant further discussion. While the IM group demonstrated significantly better mobility at discharge, with a higher proportion of patients achieving independent ambulation and fewer patients remaining bedridden, the mechanisms underlying this functional benefit are likely multifactorial.34 Reduced preoperative waiting time, lower transfusion requirements, decreased complication rates, and early rehabilitation involvement may all contribute to better functional recovery. These findings underscore the importance of considering functional outcomes alongside traditional clinical endpoints when evaluating multidisciplinary care models.
Limitations
Several limitations should be acknowledged. First, as a single-center retrospective study with non-randomized allocation, selection bias and residual confounding cannot be excluded, and the findings should be interpreted as associations rather than causal effects; important prognostic variables (frailty, ASA class, pre-fracture mobility, cognitive status, nutritional status, anemia, and anticoagulant use) were unavailable for the propensity model, making unmeasured confounding particularly relevant to the mortality finding. Second, surgical method may lie partially downstream of the admission pathway, and non-independence within matched sets was not explicitly modeled. Third, the mortality model (13 covariates with limited events) may be overfitted, and the skewed cost estimates should be interpreted with caution. Fourth, comorbidities were identified by manual keyword extraction rather than ICD coding. Fifth, although E-value analysis suggested robustness (E-value = 3.48; 1.59 for the CI upper bound), the benefits may partly reflect the concurrent geriatric pathway and green-channel organization rather than internal medicine leadership per se. Multicenter prospective randomized trials are warranted.
Despite these limitations, our data suggest that positioning internal medicine physicians as the coordinating hub of the multidisciplinary team may yield substantial clinical benefits. For hospitals with comprehensive internal medicine capabilities, establishing a dedicated geriatric hip fracture center with an emergency green channel and simultaneous admission assessment may be an effective strategy. For resource-limited settings, simplified models incorporating a designated internist conducting daily joint rounds with the orthopedic team and establishing rapid consultation response mechanisms may represent a pragmatic intermediate step toward improving perioperative outcomes.
Conclusions
This retrospective study suggests that among the study population, internal medicine-led multidisciplinary co-management may be associated with shorter preoperative waiting time and hospital stay, reduced intraoperative blood loss, allogeneic transfusion and medical costs. This model could be considered as one of the reference strategies for optimizing perioperative management of geriatric hip fracture in institutions with available multidisciplinary internal medicine resources, and further multicenter prospective studies are warranted to verify these findings.
Funding Statement
This study did not receive any specific grant from funding agencies in the public, commercial, or not-for-profit sectors.
Abbreviations
ASA, American Society of Anesthesiologists; CI, confidence interval; COPD, chronic obstructive pulmonary disease; CNY, Chinese Yuan; IM, internal medicine-led; IQR, interquartile range; MDT, multidisciplinary team; OC, orthopedic-led consultation; OR, odds ratio; PSM, propensity score matching; SMD, standardized mean difference; VTE, venous thromboembolism.
Data Sharing Statement
The data generated and analyzed during the current study are derived from the electronic medical record system of Beijing Jishuitan Hospital Guizhou Hospital. Due to patient privacy considerations, the raw data are not publicly available. De-identified datasets may be obtained from the corresponding author upon reasonable request and with approval from the institutional ethics committee.
Ethics Approval and Consent to Participate
This study was approved by the Institutional Ethics Committee of Beijing Jishuitan Hospital Guizhou Hospital (approval number: KT2026042001); although the clinical care covered by the study period preceded the ethics application, ethical approval was obtained before study-specific data collection was initiated, in line with the retrospective design. The requirement for informed consent was waived. Patient data were fully anonymized, and confidentiality was strictly maintained. All procedures were performed in accordance with the Declaration of Helsinki.
Author Contributions
All authors made a significant contribution to the work reported, whether that is in the conception, study design, execution, acquisition of data, analysis and interpretation, or in all these areas; took part in drafting, revising or critically reviewing the article; gave final approval of the version to be published; have agreed on the journal to which the article has been submitted; and agree to be accountable for all aspects of the work.
Disclosure
The authors declare that they have no competing interests.
References
- 1.Chinese Medical Association Orthopedics Branch. Guidelines for the diagnosis and treatment of osteoporotic fractures in China (2022 edition). Chin J Orthop. 2022;42(22):1473–13. Chinese. doi: 10.3760/cma.j.cn121113-20220704-00377 [DOI] [Google Scholar]
- 2.Veronese N, Maggi S. Epidemiology and social costs of hip fracture. Injury. 2018;49(8):1458–1460. doi: 10.1016/j.injury.2018.04.015 [DOI] [PubMed] [Google Scholar]
- 3.Dyer SM, Crotty M, Fairhall N, et al. A critical review of the long-term disability outcomes following hip fracture. BMC Geriatr. 2016;16(1):158. doi: 10.1186/s12877-016-0332-0 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4.Qi X, Lu L, Li H, et al. Trends, disease burden, and cause patterns of hip fracture among older adults from 1990 to 2023 in mainland China. Arch Osteoporos. 2026;21(1). doi: 10.1007/s11657-026-01695-9 [DOI] [PubMed] [Google Scholar]
- 5.Department of Medical Administration, National Health Commission of China. Guidelines for the diagnosis and management of geriatric Hip fracture (2022 edition). Beijing: National Health Commission of China; 2022. (Guo Wei Ban Yi Zheng Han [2022] No. 466). Chinese. [Google Scholar]
- 6.Roberts HJ, Rogers SE, Ward DT, et al. Protocol-based interdisciplinary co-management for hip fracture care: 3 years of experience at an academic medical center. Arch Orthop Trauma Surg. 2022;142(7):1491–1497. doi: 10.1007/s00402-020-03699-7 [DOI] [PubMed] [Google Scholar]
- 7.Bani Essa S, Anaqreh Y, Abueed M, et al. Impact of surgical timing on mortality and functional outcomes in elderly hip fracture patients: a retrospective cohort study. Acta Inform Med. 2024;32(3):196–200. doi: 10.5455/aim.2024.32.196-200 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8.Friedman SM, Mendelson DA, Bingham KW, Kates SL. Impact of a comanaged geriatric fracture center on short-term hip fracture outcomes. Arch Intern Med. 2009;169(18):1712. doi: 10.1001/archinternmed.2009.321 [DOI] [PubMed] [Google Scholar]
- 9.Rapp K, Becker C, Todd C, et al. The association between orthogeriatric co-management and mortality following hip fracture. Dtsch Arztebl Int. 2020;117(4):53–59. doi: 10.3238/arztebl.2020.0053 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10.Della Rocca GJ, Moylan KC, Crist BD, et al. Comanagement of geriatric patients with hip fractures: a retrospective, controlled, cohort study. Geriatr Orthop Surg Rehabil. 2013;4(1):10–15. doi: 10.1177/2151458513495238 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11.Liu G, Yang MH, Zhu SW, et al. Effect of orthogeriatric co-management on geriatric hip fractures in China. Orthop Surg. 2022;14(4):671–677. doi: 10.1111/os.13233 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12.Yavuz Veizi BG, Veizi E, Koutserimpas C. Geriatric care in hip fracture recovery: does it truly make a difference? A meta-analysis. Arch Orthop Trauma Surg. 2025;145(1):382. doi: 10.1007/s00402-025-05993-8 [DOI] [PubMed] [Google Scholar]
- 13.Güzel İ, Ulusoy İ, Yılmaz M, Tantekin MF, Kıvrak A. Comparison of intravenous, topical, and control groups in blood management during total hip arthroplasty: a retrospective analysis. Advances in Orthopedics. 2026;2026(1):2709548. doi: 10.1155/aort/2709548 [DOI] [Google Scholar]
- 14.Neuman MD, Silber JH, Elkassabany N, et al. Comparative effectiveness of regional versus general anesthesia for hip fracture surgery in adults. Anesthesiology. 2012;117(1):72–92. doi: 10.1097/ALN.0b013e3182545e7c [DOI] [PubMed] [Google Scholar]
- 15.Austin PC. Balance diagnostics for comparing the distribution of baseline covariates between treatment groups in propensity-score matched samples. Stat Med. 2009;28(25):3083–3107. doi: 10.1002/sim.3697 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16.Yee DK, Lau T, Fang C, et al. Orthogeriatric multidisciplinary co-management across acute and rehabilitation care improves length of stay, functional outcomes and complications in geriatric hip fracture patients: a retrospective cohort study. Hong Kong Med J. 2023;29(Suppl 2):S1–S6. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17.Salvador-Marin J, Ferrandez-Martinez FJ, Lawton CD, et al. Efficacy of a multidisciplinary care protocol for the treatment of operated hip fracture patients. Sci Rep. 2021;11:24082. doi: 10.1038/s41598-021-03415-4 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18.Gao F, Liu G, Ge Y, et al. Orthogeriatric co-managements lower early mortality in long-lived elderly hip fracture: a post-hoc analysis of a prospective study. BMC Geriatr. 2023;23:571. doi: 10.1186/s12877-023-04289-z [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19.Gao F, Wang Z, Chen Y, et al. Orthogeriatric co-management improves early outcomes in hip fractures: a post-hoc analysis of a prospective study. Clin Interv Aging. 2025;20:2231–2240. doi: 10.2147/CIA.S555649 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20.Samra T, Jain K, Kaushal V, et al. The outcome of surgically treated proximal femur fractures managed by ortho-anesthetic geriatric care pathway: a prospective observational study with 2-year follow-up. Indian J Orthop. 2023;57(6):957–966. doi: 10.1007/s43465-023-00880-z [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21.Pankratz C, Risch A, Oxen J, et al. Orthogeriatric care-outcome of different fragility fractures. Arch Orthop Trauma Surg. 2023;143(11):6641–6647. doi: 10.1007/s00402-023-04993-w [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22.Chandran M, Thiyagarajan JA, Alokail M, et al. WHO benchmarks for equitable hip-fracture care and osteoporosis treatment in older people. Nat Rev Rheumatol. 2026;22(1):62–70. doi: 10.1038/s41584-025-01319-5 [DOI] [PubMed] [Google Scholar]
- 23.Mant SJ, Amadi-Livingstone C, Ahmed MH, et al. Orthogeriatric care following hip fracture: improving post-operative outcomes in an aged population. Life. 2024;14(4):503. doi: 10.3390/life14040503 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 24.Pankratz C, Risch A, Oxen J, et al. Orthogeriatric co-management of fragility fractures: a retrospective cohort study on one-year mortality and functional outcomes across fracture types. BMC Geriatr. 2026;26:91. doi: 10.1186/s12877-025-06913-6 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 25.VanderWeele TJ, Ding P. Sensitivity analysis in observational research: introducing the E-value. Ann Intern Med. 2017;167(4):268–274. doi: 10.7326/M16-2607 [DOI] [PubMed] [Google Scholar]
- 26.Leung KL, Lau T, Chan H, et al. Orthogeriatric co-management model to improve outcome and cost-effectiveness of fragility hip fractures: abridged secondary publication. Hong Kong Med J. 2023;29(Suppl 2):S15–S17. [PubMed] [Google Scholar]
- 27.Gao Y, Zhou S, Gao W, et al. Preoperative indicators for 1-year mortality in elderly individuals following hip fracture surgery under a multidisciplinary team co-management model. Geriatr Orthop Surg Rehabil. 2025;16:21514593251356135. doi: 10.1177/21514593251356135 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 28.Xiong F, Shao XP, Rui BY. Implementation and effect evaluation of a multidisciplinary collaboration diagnosis and treatment and nursing model for traumatic geriatric Hip fracture. Chin J Modern Nurs. 2023;29(36):4956–4962. [Google Scholar]
- 29.He ZD, Yang W, Pang ZL. Clinical efficacy evaluation of multidisciplinary collaboration model in very elderly hip fracture patients. J Med Theor Pract (Yixue Lilun yu Shijian). 2024;37(11). Chinese. doi: 10.3969/j.issn.1001-7584.2024.11.012 [DOI] [Google Scholar]
- 30.Piao HW, Zhang HT, Ma QS, et al. Effect of enhanced recovery after surgery combined with multidisciplinary collaboration on perioperative indicators in elderly hip fracture patients. Clin Med Res Pract (Linchuang Yixue Yanjiu yu Shijian). 2023;8(15). Chinese. doi: 10.3969/j.issn.1674-4721.2023.15.015 [DOI] [Google Scholar]
- 31.Xu M, Zhong MS, Liu M, et al. Feasibility study of multidisciplinary collaboration management model in total hip arthroplasty in elderly patients. Chin J Bone Joint (Zhongguo Gu Yu Guanjie Zazhi). 2024;13(10):854–859. Chinese. [Google Scholar]
- 32.Wang ZQ. Effect of MDT-based integrated medical-nursing information nursing model in elderly hip fracture. J Med Theor Pract (Yixue Lilun yu Shijian). 2023;36(10):1746–1748. Chinese. doi: 10.19381/j.issn.1001-7585.2023.10.048 [DOI] [Google Scholar]
- 33.Xie T, Rui YF, Liu SQ, et al. Preliminary clinical efficacy of establishing a rapid ICU admission pathway for elderly Hip fracture patients under the enhanced recovery after surgery concept. Natl Med J China (Zhonghua Yi Xue Za Zhi). 2020;100(37). Chinese. [Google Scholar]
- 34.Lari A, Haidar A, AlRumaidhi Y, et al. Predictors of mortality and length of stay after hip fractures - a multicenter retrospective analysis. J Clin Orthop Trauma. 2022;28:101853. doi: 10.1016/j.jcot.2022.101853 [DOI] [PMC free article] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Data Availability Statement
The data generated and analyzed during the current study are derived from the electronic medical record system of Beijing Jishuitan Hospital Guizhou Hospital. Due to patient privacy considerations, the raw data are not publicly available. De-identified datasets may be obtained from the corresponding author upon reasonable request and with approval from the institutional ethics committee.